2.1. LYSOSOMAL STORAGE DISEASES
Lysosomes, which are present in all animal cells, are acidic cytoplasmic organelles that contain an assortment of hydrolytic enzymes. These enzymes function in the degradation of internalized and endogenous macromolecular substrates. When there is a lysosomal enzyme deficiency, the deficient enzyme's undegraded substrates gradually accumulate within the lysosomes causing a progressive increase in the size and number of these organelles within the cell. This accumulation within the cell eventually leads to malfunction of the organ and to the gross pathology of a lysosomal storage disease, with the particular disease depending on the particular enzyme deficiency. More than thirty distinct, inherited lysosomal storage diseases have been characterized in humans.
A few examples of lysosomal storage diseases (and their associated deficient enzymes) include Fabry disease (.alpha.-galactosidase), Farber disease (ceramidase), Gaucher disease (glucocerebrosidase), G.sub.ml gangliosidosis (.beta.-galactosidase), Tay-Sachs disease (.beta.-hexosaminidase), Niemann-Pick disease (sphingomyelinase), Schindler disease (.alpha.-N-acetylgalactosaminidase), Hunter syndrome (iduronate-2-sulfatase), Sly syndrome (.beta.-glucuronidase), Hurler and Hurler/Scheie syndromes (iduronidase), and I-Cell/San Filipo syndrome (mannose 6-phosphate transporter).
One proven treatment for lysosomal storage diseases is enzyme replacement therapy in which an active form of the enzyme is administered directly to the patient. However, abundant, inexpensive and safe supplies of therapeutic lysosomal enzymes are not commercially available for the treatment of any of the lysosomal storage diseases.
2.1.1. GAUCHER DISEASE AND TREATMENT
Gaucher disease is the most common lysosomal storage disease in humans, with the highest frequency encountered in the Ashkenazi Jewish population. About 5,000 to 10,000 people in the United States are afflicted with this disease (Grabowski, 1993, Adv. Hum. Genet. 21:377-441). Gaucher disease results from a deficiency in glucocerebrosidase (hGC; glucosylceramidase; acid .beta.-glucosidase; EC 3.2.1.45). This deficiency leads to an accumulation of the enzyme's substrate, glucocerebroside, in reticuloendothelial cells of the bone marrow, spleen and liver, resulting in significant skeletal complications such as bone marrow expansion and bone deterioration, and also hypersplenism, hepatomegaly, thrombocytopenia, anemia and lung complications (Grabowski, 1993, supra; Lee, 1982, Prog. Clin. Biol. Res. 95:177-217; Brady et al., 1965, Biochem. Biophys. Res. Comm. 18:221-225).
hGC replacement therapy has revolutionized the medical care and management of Gaucher disease, leading to significant improvement in the quality of life of many Gaucher patients (Pastores et al., 1993, Blood 82:408-416; Fallet et al., 1992, Pediatr. Res. 31:496-502). Studies have shown that regular, intravenous administration of specifically modified hGC (Ceredase.TM., Genzyme Corp.) can result in dramatic improvements and even reversals in the hepatic, splenic and hematologic manifestations of the disease (Pastores et al., 1993, supra; Fallet: et al., 1992, supra; Figueroa et al., 1992, N. Eng. J. Med 327:1632-1636; Barton et al., 1991, N. Eng. J. Med. 324:1464-1470; Beutler et al., 1991, Blood 78:1183-1189). Improvements in associated skeletal and lung complications are possible, but require larger doses of enzyme over longer periods of time.
Despite the benefits of hGC replacement therapy, the source and high cost of the enzyme seriously restricts its availability. Until recently, the only commercial source of purified hGC has been from pooled human placentae, where ten to twenty kilograms (kg) of placentae yield only 1 milligram (mg) of enzyme. From five hundred to two thousand kilograms of placenta (equivalent to 2,000-8,000 placentae) are required to treat each patient every two weeks. Current costs for HGC replacement therapy range from $55 to $220/kg patient body weight every two weeks, or from $70,000 to $300,000/year for a 50 kg patient. Since the need for therapy essentially lasts for the duration of a patient's life, costs for the enzyme alone may exceed $15,000,000 during 30 to 70 years of therapy.
A second major problem associated with treating Gaucher patients with glucocerebrosidase isolated from human tissue (and perhaps even from other animal tissues) is the risk of exposing patients to infectious agents which may be present in the pooled placentae, e.g., human immuno-deficiency virus (HIV), hepatitis viruses, and others.
Accordingly, a new source of hGC is needed to effectively reduce the cost of treatment and to eliminate the risk of exposing Gaucher patients to infectious agents.
2.1.2. HURLER SYNDROME AND TREATMENT
Hurler syndrome is the most common of the group of human lysosomal storage disorders known as the mucopolysaccharidoses (MPS) involving an inability to degrade dermatan sulfate and heparan sulfate. Hurler patients are deficient in the lysosomal enzyme, .alpha.-L-iduronidase (IDUA), and the resulting accumulation of glucosaminoglycans in the lysosomes of affected cells leads to a variety of clinical manifestations (Neufeld & Ashwell, 1980, The Biochemistry of Glycoproteins and Proteoglycans, ed. W. J. Lennarz, Plenum Press, N.Y.; pp. 241-266) including developmental delay, enlargement of the liver and spleen, skeletal abnormalities, mental retardation, coarsened facial features, corneal clouding, and respiratory and cardiovascular involvement. Hurler/Scheie syndrome (MPS I H/S) and Scheie syndrome (MPS IS) represent less severe forms of the disorder but also involve deficiencies in IDUA. Molecular studies on the genes and cDNAs of MPS I patients has led to an emerging understanding of genotype and clinical phenotype (Scott et al., 1990, Am. J. Hum. Genet. 47:802-807). In addition, both a canine and feline form of MPS I have been characterized (Haskins et al., 1979, Pediat. Res. 13:1294-1297; Haskins and Kakkis, 1995, Am. J. Hum. Genet. 57:A39 Abstr. 194; Shull et al., 1994, Proc. Natl. Acad. Sci. USA, 91:12937-12941) providing an effective in vivo model for testing therapeutic approaches.
The efficacy of enzyme replacement in the canine model of Hurler syndrome using human IDUA generated in CHO cells was recently reported (Kakkis et al., 1995, Am. J. Hum. Genet. 57:A39 (Abstr.); Shull et al., 1994, supra). Weekly doses of approximately 1 mg administered over a period of 3 months resulted in normal levels of the enzyme in liver and spleen, lower but significant levels in kidney and Lungs and very low levels in brain, heart, cartilage and cornea (Shull et al., 1994, supra. Tissue examinations showed normalization of lysosomal storage in the liver, spleen and kidney, but no improvement in heart, brain and corneal tissues. One dog was maintained on treatment for 13 months and was clearly more active with improvement in skeletal deformities, joint stiffness, corneal clouding and weight gain (Kakkis et al., 1995, supra. A single higher-dose experiment was quite promising and showed detectable IDUA activity in the brain and cartilage in addition to tissues which previously showed activity at the lower does. Additional higher-dose experiments and trials involving longer administration are currently limited by availability of recombinant enzyme. These experiments underscore the potential of replacement therapy for Hurler patients and the severe constraints on both canine and human trials due to limitations in recombinant enzyme production using current technologies.
2.2. BIOSYNTHESIS OF LYSOSOMAL ENZYMES
Soluble lysosomal enzymes share initial steps of biosynthesis with secretory proteins, i.e., synthesis on the ribosome, binding of the N-terminal signal peptide to the surface of the rough endoplasmic reticulum (ER), transport into the lumen of the ER where the signal peptide is cleaved, and addition of oligosaccharides to specific asparagine residues (N-linked), followed by further modifications of the nascent protein in the Golgi apparatus (von Figura and Hasilik, 1986, Annu. Rev. Biochem. 55:167-193). The N-linked oligosaccharides can be complex, diverse and heterogeneous, and may contain high-mannose residues. The proteins undergo further processing in a post-ER, pre-Golgi compartment and in the cis-Golgi to form either an N-linked mannose 6-phosphate (M-6-P) oligosaccharide-dependent or N-linked M-6-P oligosaccharide-independent recognition signal for lysosomal localized enzymes (Kornfeld & Mellman, 1989, Ann. Rev. Cell Biol., 5:483-525; Kaplan et al., 1977, Proc. Natl. Acad. Sci. USA 74:2026). The presence of the M-6-P recognition signal results in the binding of the enzyme to M-6-P receptors (MPR). These bound enzymes remain in the cell, are eventually packaged into lysosomes, and are thus segregated from proteins targeted for secretion or to the plasma membrane.
Although many lysosomal enzymes are soluble and are transported to lysosomes by MPRs, integral membrane and membrane-associated proteins (notably hGC) are targeted and transported to lysosomes independent of the M-6-P/MPR system (Kornfeld & Mellman, 1989, Erickson et al., 1985). hGC does not become soluble after translation, but instead becomes associated with the lysosomal membrane by means which have not been elucidated (von Figura & Hasilik, 1986, Annu. Rev. Biochem. 55:167-193; Kornfeld and Mellman, 1989, Annu. Rev. Cell Biol. 5:483-525).
hGC is synthesized as a single polypeptide (58 kDa) with a signal sequence (2 kDa) at the amino terminus. The signal sequence is co-translationally cleaved and the enzyme is glycosylated with a heterogeneous group of both complex and high-mannose oligosaccharides to form a precursor. The glycans are predominately involved in protein conformation. The "high mannose" precursor, which has a molecular weight of 63 Kda, is post-translationally processed in the Golgi to a 66 Kda intermediate, which is then further modified in the lysosome to the mature enzyme having a molecular weight of 59 Kda (Jonsson et al., 1987, Eur. J. Biochem. 164:171; Erickson et al., 1985, J. Biol. Chem., 260:14319).
The mature hGC polypeptide is composed of 497 amino acids and contains five N-glycosylation amino acid consensus sequences (Asn-X-Ser/Thr). Four of these sites are normally glycosylated. Glycosylation of the first site is essential for the production of active protein. Both high-mannose and complex oligosaccharide chains have been identified (Berg-Fussman et al., 1993, J. Biol. Chem. 268:14861-14866). hGC from placenta contains 7% carbohydrate, 20% of which is of the high-mannose type (Grace & Grabowski, 1990, Biochem. Biophys. Res. Comm. 168:771-777). Treatment of placental hGC with neuraminidase (yielding an asialo enzyme) results in increased clearance and uptake rates by rat liver cells with a concomitant increase in hepatic enzymatic activity (Furbish et al., 1981, Biochim. Biophys. Acta 673:425-434). This glycan-modified placental hGC is currently used as a therapeutic agent in the treatment of Gaucher's disease. Biochemical and site-directed mutagenesis studies have provided an initial map of regions and residues important to folding, activator interaction, and active site location (Grace et al., 1994, J. Biol. Chem. 269:2283-2291).
The complete complementary DNA (cDNA) sequence for hGC has been published (Tsuji et al., 1986, J. Biol. Chem. 261:50-53; Sorge et al., 1985, Proc. Natl. Acad. Sci. USA 82:7289-7293), and E. coil containing the hGC cDNA sequence cloned from fibroblast cells, as described (Sorge et al., 1985, supra), is available from the American Type Culture Collection (ATCC) (Accession No. 65696).
Recombinant methodologies have the potential to provide a safer and less expensive source of lysosomal enzymes for replacement therapy. However, production of active enzymes, e.g., hGC, in a heterologous system requires correct targeting to the ER, and appropriate N-linked glycosylation at levels or efficiencies that avoid ER-based degradation or aggregation. Since mature lysosomal enzymes must be glycosylated to be active, bacterial systems cannot be used. For example, hGC expressed in E. coli is enzymatically inactive (Grace & Grabowski, 1990, supra).
Active monomers of hGC have been purified from insect cells (Sf9 cells) and Chinese hamster ovary (CHO) cells infected or transfected, respectively, with hGC cDNA (Grace & Grabowski, 1990, supra; Grabowski et al., 1989, Enzyme 41:131-142). A method for producing recombinant hGC in CHO cell cultures and in insect cell cultures was recently disclosed in U.S. Pat. No. 5,236,838. Recombinant hGC produced in these heterologous systems had an apparent molecular weight ranging from 64 to 73 kDa and contained from 5 to 15% carbohydrate (Grace & Grabowski, 1990, supra; Grace et al., 1990, J. Biol. Chem. 265:6827-6835). These recombinant hGCs had kinetic properties identical to the natural enzyme isolated from human placentae, as based on analyses using a series of substrate and transition state analogues, negatively-charged lipid activators, protein activators (saposin C), and mechanism-based covalent inhibitors (Grace et al., 1994, supra; Berg-Fussman et al., 1993, supra; Grace et al., 1990, J. Biol. Chem. 265:6827-6835; Grabowski et al., 1989, supra). However, both insect cells and CHO cells retained most of the enzyme rather than secreting it into the medium, significantly increasing the difficulty and cost of harvesting the pure enzyme (Grabowski et al., 1989, supra).
Accordingly, a recombinant system is needed that can produce human or animal lysosomal enzymes in an active form at lower cost, and that will be appropriately targeted for ease of recovery.
2.3. MAMMALIAN LYSOSOMES VERSUS PLANT VACUOLES
Because plants are eukaryotes, plant expression systems have advantages over prokaryotic expression systems, particularly with respect to correct processing of eukaryotic gene products. However, unlike animal cells, plant cells do not possess lysosomes. Although the plant vacuole appears functionally analogous to the lysosome, plants do not contain MPRs (Chrispeels, 1991, Ann. Rev. Pl. Phys. Pl. Mol. Biol. 42:21-53; Chrispeels and Tague, 1991, Intl. Rev. Cytol. 125:1-45), and the mechanisms of vacuolar targeting can differ significantly from those of lysosomal targeting. For example, the predominant mechanism of vacuolar targeting in plants does not appear to be glycan-dependent, but appears to be based instead on C- or N-terminal peptide sequences (Gomez & Chrispeels, 1993, Plant Cell 5:1113-1124; Chrispeels & Raikhal, 1992, Cell 68:613-618; Holwerda et al., 1992, Plant Cell 4:307-318; Neuhaus et al., 1991, Proc. Natl. Acad. Sci. USA 88:10362-10366; Chrispeels, 1991, supra; Chrispeels & Tague, 1991, supra; Holwerda et al., 1990, Plant Cell 2:1091-1106; Voelker et al., 1989, Plant Cell 1:95-104). As a result, plants have not been viewed as appropriate expression systems for lysosomal enzymes which must be appropriately processed to produce an active product.